Immediate Assistance Required: Ongoing Service Disruption and Request for Compensation
Dear RunPod Customer Support Team,
I am currently experiencing a critical issue with your GPU instance services, which demands immediate attention. The download speeds have drastically reduced, significantly affecting my workflow. Typically, downloading 10 LohA files takes mere seconds, but it's now taking over 20 minutes. Moreover, I have an additional 15 LokR files pending download, which, at this rate, will take several hours – a situation that is unacceptable and highly inefficient.
Regrettably, my attempts to reach out to your support team have been met with silence. This lack of response is concerning, especially given the ongoing nature of the problem and its continuous drain on my credits, effectively leading to a financial loss.
Given these circumstances, I am requesting:
Immediate intervention and technical support to rectify the download speed issue.
An explanation for the sudden decrease in service performance.
Compensation for the credits expended and the significant delay caused, considering the urgency and impact of this issue.
The resolution of this matter is critical. This ongoing disruption not only impedes my current projects but also casts doubt on the dependability of RunPod's services. I am looking forward to a prompt response and a just solution to compensate for the inconvenience and losses I am currently facing.
Thank you for your immediate attention to this pressing issue.
12 Replies
Hey, really sorry about this. Can you confirm the details of your deployment, specifically if you're in a specific region or datacenter and where you're downloading from (us east s3 for example)? Also if you're using serverless or gpu instances. We will dig into this immediately.
it is in US
bm8e1db676gnkv
This is the pod id
It's a single community cloud instance? We'll get it checked out but it's difficult to guarantee bandwidth on those types of instances. Has this been consistent for a long period of time or is it intermittent?
I just run it temporary for some training
Would be ok if there's an option for me to change the instance type but keep the files
but nothing I can do
I restarted it several times the problem still remain
and my credits is still burning
for nothing
Are you downloading to or from the instance?
from
What method are you using?
just download
from the jupyter
http protocol
Gotcha, something you can try is to use runpodctl
It may be that you got load balanced to a proxy that is getting hammered
Since the http jupyter needs to use a proxy web server to facilitate your download
There's a section on file transfers in the docs. Direct tcp can also be used
it's not very convinient, can it download multiple files with a single command?
Yeah if you send a directory then it'll zip it and send it by default
ok this way is much more faster